J. Org. Chem. 1993,58, 4851-4854
4851
lSC NMR Spectroscopic Study of Para-Substituent Effects in Highly Crowded 1,l'-Diadamantylbenzyl Cations' George A. Olah,' Michael D. Heagy, and G. K. Surya Prakash' The Donald P. and Katherine B. Loker Hydrocarbon Research Institute and Department of Chemistry, University of Southern California, University Park, Los Angeles, California 90089-1661 Received April 5, 1993
A series of para-substituted 1,l'-diadamantylbenzyl cations were studied under stable ion conditions at low temperatures in superacid solutions. These studies were informative in probing the effects of ring substituents and their ability to delocalize positive charge from the carbenium center in highly crowded systems. The observed 13CNMR chemical shifts of the carbocation center in comparison with those of the correspondingcumyl cations indicate the effects of steric crowdingof the diadamantyl substituents in minimizing p-?r interactions. This renders the carbocationic l3C NMR chemical shifts virtually unaffected by para-substituents.
Introduction
Scheme I
The Gassman-Fentiman tool of increasing electron demand2 has been frequently employed in probing the substituent-inducedelectroniceffects of electron-deficient systems using NMR spectroscopy. The technique involves systematic variation of the aryl ring substituents (varying from electron-releasingto electron-withdrawing)adjacent to a developing carbenium center in an effort to detect the onset of other electronic interactions such as u,TU, and u delocalization in the system. Although the "tool of increasing electron demand" applied to stable carbocations is more sensitive in detecting electronic interactions than the usual solvolysisstudies wherein solvation significantly masks the electron demand of the cation center in the solvolytic transition state, the method has limitations. A phenyl group even with electron-withdrawingsubstituents can still delocalize positive charge in ita u-system. This sometimes makes the method insensitive to detect subtle interactions which are operative, when the aryl group is replaced by a hydrogen substituent. Our interest focused on a system where p u delocalization is effectivelyscreened from the carbocation center. Earlier workaestablishedthe tertiary a,a-di(1-adamanty1)phenylmethyl cation as showing interesting spectroscopic behavior since little charge is delocalized to the phenyl ring. The chemical shift comparison of cations 1 and 2 emphasizes the difference in charge distribution which result from very unfavorable steric interactions between the aryl ring and adamantyl @hydrogens. The ring is forced to twist and largely eliminates conjugation with the empty p-lobe of the cationic center. This exceptional exclusion of resonance effects in association with carbocationic intermediates with adjacent aromatic systems
FS0,W S0,CIF
(1) Stable Carbocations 288. For part 287 see Olah,G.A.; Liao, Q.; Casanova, J.; Bau, R.; Prakash, G. K. S. J. Am. Chem SOC.,in press. (2) (a) Prakaeh, G.K. 9.; Iyer, P. S. Reo. Chem. Intermed. 1988, 9, 1969, 65-116. (b)Gaesman! P. G.; Fentiman, A. F.,Jr. J. Am. Chem. SOC. 91,1545-1548. (c) Richey, H. G., Jr.; Nichols, J. D.; Gamman, P. G.; Fentiman, A. F., Jr.; Winstein,S.; Brookhart, M.; Lustgarten, R. K. J. Am. Chem. SOC.1970,92,3783-3784. (3) Olah,G. A.; Prakash, G. K. S.; Liang, G. L.; Schleyer, P. v. R.; Graham, W. D. J. Org. Chem. 1982, 47, 1040-1047. Tanida, H.; Mataumura, H. J. Am. Chem. SOC.1973,95,1586-1593. Duieman, W.; Ruchardt,C. Chem. Ber. 1973,106,1083-1098. Lomas, J. S.; Dubis, J. E.;J. Org.Chem. 1975,40,3303-3304. Tetrahedron 1978,34,1597-1604.
I
2
2
3
4
prompted us to investigatethis unique system for inductive effects only.
2-H
1
Consequently, we prepared a series of related parasubstituted 1,l'-diadamantylbenzyl cations and studied them by 13CNMR spectroscopy. The needed precursors, para-substituted 1,l'diadamantylbenzyl alcohols suitable for studies of increasing electron demand were prepared from 1,l'-diadamantyl ketone* with aryllithiums? Ionization of the alcohol in FS03H/SO&lF at -78 "C generated the benzylcation as a dark red solution (Scheme
I). Results and Discussion The 13CNMR spectrum (Figure 1)of 1,l'-diadamantylp-tolyl cation shows a highly deshielded benzylic cationic center at 13C 6 286 as compared to the cumyl cation resonance at 255 ppm.21s Such extraordinary deshielding (4) Olah,G.A.; Farm, 0.; Prakash, G. K. S. Synthesie 1985,11401142. (5) Choi, C. A.; Pinkerton,A. Fry, J. L. J. Chem. SOC.Chem. Commun. 1987,225. Olah,G.A.; Prakash, G.K. S.; Liang, G.L.; Schleyer, P. v. R.; Graham, W. D. J. Org. Chem. 1982,47,1040-1047. (6) Olah,G.A.; Porter, R. D.; Juell, C. L.; White,A.M. J. Am. Chem. Soc. 1972,94,2044-2052. Olah,G. A.; Berrier, A. L.; Prakash, G.K. S. h o c . Natl. Acad. Sci. U.S.A.1988, 78,1988-2002.
0022-326319311958-4851$04.00/0 0 1993 American Chemical Society
4862
J. Org. Chem., Vol. 58,No.18,1993
Olah et al.
Table I. W NMR Spectroscopic Data. on Para-Substituted Diadamantylbenzyl Cations in SO&lF at -78 O C
4-2 OCH3 CH3 F H CF3
C+
a
B
Y
6
283.0 286.1 284.7 286.5 287.1
67.2 66.7 67.6 65.9 68.5
39.6 39.7 39.7 38.3 39.8
29.2 29.5 29.5 26.4 27.9
33.7 33.8 33.7 32.5 33.7
ipso 135.8 145.4 132.9 136.2 140.7
m 112.9 125.6 115.lC 122.8 123.W
0
130.3 127.7 128.3b 125.9 124.8
P 157.7 134.7 164.6d 131.4 133.lf
C+(UYlCP) 237 260 265 271 2Mh
2 57.6 20.6 121.38
a Carbon-13 chemical shifta are in parts per million from the external acetone (capillary) signal. (‘JCp) = 3.2 Hz. ( 2 J c ~ =)22.4 Hz. (~JGF) = 257.4 Hz.e ( 3 J ~= ~11.9 ) Hz. f (%JGF) = 33.4 Hz. g WGF)= 256.6 Hz. 13C NMR chemical shifta of 1-aryl-1-cyclopentylcations are included for comparison.
c, C,4 280
Figure 1. NMR spectrum (75.4 MHz, proton decoupled) of 1,l’-diadamantyl-p-tolyl carbocation. Absorptions due to acetone-& are indicated by asterisks.
230
4
4 -3
/
OCHJ
/
-2
oc*
.‘
0
Figure 3. Chemical shift correlation plot of 1-aryl-1-cyclopentyl and 1,l’-diadamantylbenzyl cation chemical shifts versus oc+ substituent constants.
Figure 2. 13C NMR spectrum (75.4 MHz, proton decoupled) of 1,l’-diadamantyl-p-tolylmethanol.
can be explained only by the decreased (if not nonexistent p-r overlap) with the carbocationic center. This observation also indicates that the two bulky adamantyl groups prohibit coplanar ring rotation of the phenyl group into conjugation with the empty p-orbital of the cationic center. MMX calculations of the cationic species complement this interpretion via high rotational energy barriers as the phenyl ring rotates through its molecular axis.? The I3C resonances of the ortho and meta carbons shown in Figure 1indicate one peak each, which we attribute to the adopted ground state CzVsymmetry of the cation. However, full 360° rotation of the aryl group cannot be excluded. The observation at room temperature of diastereotopic 13C aromatic resonances (Figure 2) from the parent neutral alcohol (C,symmetry) was interpreted as a “freezing” of the phenyl ring about the Cipso-C+ bond into a fixed position, suggesting strong steric interactions imposed by the adamantyl substituents. MMX Molecular modeling calculations were carried out using PCModelversion 4.0 Serena Software. Singlepoint calculationswithout optimization were carried out at IOo incrementsof rotation about-,iC C+ bond. (7)
The 13CNMR data for the 1,l’-diadamantylbenzyl ions are summarized in Table I. We have also included the 13C NMR chemical shifts of the cationic center of model 1-aryl1-cyclopentyl cations (a system similar to cumyl)* for comparison. The range of 13Cchemical shifts for the 1,l’diadamantylbenzyl cationic centers varies by only f 2 ppm throughout the series. The aromatic carbons remain relatively unaffected by the charge as evidenced by their 13C NMR resonances in comparison to parent neutral compounds. Concomitantly, we observe up to ca. 23 ppm downfield chemical shift of the C, adamantyl carbons which indicates extensive hyperconjugative interactions that compensate for the lack of resonance stabilization from the aryl group. A plot of the carbocationic chemical shifts of 1,l’diadamantylbenzyl cations and 1-aryl-1-cyclopentyl cations versus &+indicates a near unity slope for l-aryl-lcyclopentyl cations with excellent correlation (Figure 3). However, these two systems appear to have considerably different carbocationic environments as indicated by their slopes. The 1,l’-diadamantylbenzyl cations appear to have almost zero slope in comparison to that of the l-aryl-lcyclopentyl system. Removal of n-resonance interaction with the adjacent carbocationic center in these highly crowded systems plays a significant role in dampening the effects of substituents ‘Z’ at the para position. Not surprisingly, the inductive effects of the various substituents are negligible since the interaction of bond dipoles is transmitted through the bonds by successive polariza(8) c++ substituent constantswere basedon the cationicchemicalshifts of substituted cumylcationspreviously developed by Brown et al.9 These constants taken into account the increased electron demand of carbocations in superacidmedium. 1-Aryl-1-cyclopentylcations provide a model set of ”ordinary”cations for comparison.
13C NMR
Study of Crowded 1,l'-Diadamantylbenzyl Cations
Table 11. Carbocation Charge-Densities Data Based on Semiempirical PM3 Calculations. 2 benzyl 1-aryl-1-cyclopentyl 1,l'-diadamantylbenzyl 0.265 0.422 OCH3 0.200 0.303 0.471 0.250 CH3 0.317 0.485 F 0.271 0.325 0.489 H 0.285 0.352 0.509 0.325 CF3 a Calculations were performed with the Spartan program on an IBM RS/6000 Model 560 computer.
J. Org. Chem., Vol. 58, No. 18, 1993 4853 3. A possible explanation for this discrepancy may be due to the inability of semiempiricalcalculations (such as PM3)
to handle carbocationic structures with shallow energy minima. In summary, the para-substituted 1,l'-diadamantylbenzyl cation system shows a remarkably deshielded carbocationic center for benzylic cations in solution. This extraordinary deshielding is the result of high steric strain associated with the 1,l'-diadamantyl framework. The prohibitively high rotational barriers prevent complete p-?r resonance overlap with the adjacent empty p-orbital. T o compensate for the lack of resonance stabilization, significant hyperconjugative interactions via the adamantyl substituents are observed from the I3C NMR chemical shifts. The subtle chemical shift differences of the cationic center are virtually independent of substituent effects in the para position. Experimental Section
6 03
0.2! 0.20
.
. . .
I
0 25
. . .
.
I
,
r
. . I
0.30
0 35
Q-Values of p - Substituted benzyl cations
Figure 4. PM3 charge density correlation plot of carbocationic centers for 1,l'-diadamantylbenzyl cations and arylcyclopentyl versus benzyl cations.
tion. Thus, inductive effects drop off rapidly as the cationic site is four bonds removed from the substituent. T h e subtle differences in chemical shift shown in Table I may arise from solvation effects of the carbocationic center depending on slight differences in acid concentration of the sample. However, although lesser in magnitude, the positive slope follows the same trend as the l-aryl-lcyclopentyl system as well as other substituted aromatic systems adjacent t o a cationic center. T o coordinate the NMR study with molecular modeling, theoretical calculations on the studied carbocationic systems were carried out at the PM3 level.'O Charge densities (&-values shown in Table 11) at the carbocationic centers were calculated for para-substituted 1,l'-diadamantylbenzyl, 1-aryl-1-cyclopentyl, as well as para-substituted benzyl cations. The calculated &-values of both 1,l'-diadamantylbenzyl and 1-aryl-1-cyclopentyl cations were plotted against those of the sterically unencumbered benzyl cations (see Figure 4). The plots from Figure 4 clearly indicate that more positive charge is localized in the sterically congested 1,l'-diadamantylbenzyl system (as reflected by higher &-values) compared to l-aryl-lcyclopentyl cations. These observations support our experimental findings that little positive charge is dispersed into the aryl ring due t o steric inhibition of resonance posed by the two bulky adamantyl groups. However, the nearly identical slopes shown in Figure 4 is misleading for both systems and does not corroborate with t h e 13C chemical shift/&+ correlations reported in Figure (9) Brown, H.C.; Kelly, D. P.; Periaeamy, M. R o c . Natl. Acad. Sci. U S A . 1980,77,6966-62. (10) Calculations were performed with the Spartan program on an IBM REV6000 Model 560 computer.
THF and diethyl ether were distilled from sodium-benzophenone immediately before use. Lithium metal dispersion in mineral oil was obtained from Aldrich and used as received. 70eV Mass spectra were recorded on a Finnigan-Mat/Incos-50mass spectrometer,using DEP probe. Melting points were determined using a Mettler FP1 melting point apparatus. lH and "W NMR spectra were recorded on a Varian VXR-200 or Varian UNITY300 equipped with a variable temperature probe. The lH and I3C NMR chemical shifts for the carbocations were referenced with respect to the external capillary &-acetone. 1,l'-Diadamantylphenylmethanol (3a). Under an argon atmosphere in flame-driedglassware, a solution of chlorobenzene (0.85 mL, 8 mmol) and (0.5 g, 1.7 mmol) 1,l'-diadamantyl ketone in THF (10 mL) was added dropwise at 0 OC with vigorous stirring to a mixture of lithium dispersion (21 mmol, 3.3% Na) in THF over a 5-min period. Following overnight stirring at room temperature, the excess lithium was destroyed by careful addition of saturated aqueous ammonium chloride solution, and the mixture was extracted with pentane. The pentane extract was then washed with concd brine, dried (Na2SOJ, and concentrated in uacuo. Purification by elution over a silicagel column (hexaneether, 60:lO) followed by recrystallization from pentane-ether (9010) gave pure alcohol 0.233 g (46%): mp 212.4 OC; IR (CC4) v 3348, 2908, 2840, 1549, 1250, 1006, 825 cm-l; lH NMR (300 MHz, CDCls) 6 7.5-6.8, 1.98, 1.6; lsC NMR (75 MHz, CDCU 6 144.1,128.4,127.9,127.3,125.7,125.3,83.1,44.7,39.4,36.9,29.2. Anal. Calcd for C ~ H M OC, : 86.12; H, 9.70. Found: C, 85.89; H, 9-90. 1,l'-Diadamantyl-ptolylmethanol(3b). Preparation similar to 1,l'-diadamantylphenylmethanol (3a). A solution of p-chlorotoluene(1.01 mL, 8.4 mmol) and 1,l'-diadamantyl ketone (0.5 g, 1.7 mmol) in THF (10 mL) was added dropwise at 0 OC with vigorous stirring to a mixture of lithium dispersion(21 mmol, 3.3% Na) in THF over a 5-min period. Following overnight stirring at room temperature, the excess lithium was destroyed by careful addition of saturated aqueous ammonium chloride solution, and the mixture was extracted with pentane. The pentane extract was then washed with concd brine, dried (NazSO,), and concentrated in uacuo. Purification by elution over silica gel column (hexane-ether, 6010) followed by recrystallization from pentane-ether (9010) gave pure alcohol 0.185 g (38%): mp 274.7 OC; IR (CCb) v 3348, 2908, 2840, 1549,1256, 1006,780 cm-1; 1H NMR (300 MHz, CDCW 6 7.6-7.0,2.1,1.9,1.6; "CNMR (75MHz, CDCl3) b 140.8,135,128.2,127.9,127.7,126.3, 83.1, 44.7, 39.4, 37.0, 29.4, 20.8. Anal. Calcd for C&& C, 86.09; H, 9.81. Found C, 86.13; H, 9.76. l,lf-Diadamantyl-panisy1methanol(3c). Under an argon atmosphere in flame-driedglassware,a solution ofp-chloroanisole (2.38mL, 17mmol)and 1,l'-diadamantylketone (0.5g, 1.7mmol) in THF (10mL) was added dropwise at 0 "C with vigorous stirring to a mixture of lithium dispersion (21 mmol, 3.3% Na) in THF over a 5-min period. The temperature was maintained at constant ice-bath temperature for 5 h and the excess lithium was destroyed
4854 J. Org. Chem., Vol. 58, No.18, 1993 by careful addition of saturated aqueous ammonium chloride solution. The mixture was extracted with ether. The ether extract was then washed with concd brine, dried (NaSO,), and concentrated in uacuo. Purification by elution over silica gel column (hexane-ethyl acetate, %lo) followed by recrystallization from boiling ethanol yielded 0.300 g (60% of the alcohol: mp 294 OC; IR (CCL) Y 3348,2908,2840,1549,1250,1006,780 cm-l; 1H NMR (300 MHz, CDCb) 6 7.6-6.8,3.5,1.9,1.6;13CNMR (75 MHz, CDCb) 6 158.7,137.4, 130.3,129.9,112.5,111.8,83.4,55.2, 45.6, 40.2, 38.1, 30.5. Anal. Calcd for CZSHSSOZ:C, 82.71; H, 9.42. Found C, 82.29; H, 9.64. l,l'-Diadamantyl-pfluorophenylmethanol(3d). Under an argon atmosphere in flame-dried glassware,a solution of l-bromo4-fluorobenzene (2.08 mL, 17 "01) and 1,l'-diadamantyl ketone (0.5 g, 1.7 mmol) in diethyl ether (10 mL) was added dropwise at 0 OC with vigorous stirring to a mixture of lithium dispersion (21 mmol, 3.3% Na) in ether over a 5-min period. The reaction was allowed to warm to room temperature, stirred vigorously for 5 days, and tracked by TLC analysis. The excess lithium was destroyed by careful addition of saturated aqueous ammonium chloride solution and the mixture was extracted with ether. Purification by elution through silica gel column (hexane-ethyl acetate, %lo) followed by recrystallization from boiling cyclohexane yielded 0.15g (32%)of the alcohol: mp 213 "C; IR (CC4) Y 3348, 2908, 2840, 1540, 1280, 1069, 785 cm-l; lH NMR (300 MHz, CDCls) 6 7.8-6.9, 1.9, 1.6; '9c NMR (75 MHz, CDCls) 6 161.2 (~JGF = 243.6 HZ), 139.3,130.7 ('JCp = 7.2 HZ), 128.5 ('JCp = 7.2 Hz), 114.4 ( V c p = 20.7 Hz), 111.6 ( z J =~20.7 ~ Hz), 83, 44.2,1,39.4,37.1,29.5;'9F NMR (188 MHz, CDCls) 6 -118. Anal. Calcd for CnH&O C, 82.19; H, 8.94; F, 4.81. Found C, 81.73; H, 8.50; F, 4.57.
l,l'-Diadamantyl-p(trifluoromethyl)phenylmetlunol(38). Under an argon atmosphere in flame-dried glassware,a solution
Olah et al. of l-bromo-4-(fluoromethyl)benzene(2.7 mL, 17 mmol) and 1,l'diadamantylketone (0.5 g, 1.7 mmol) in diethyl ether (10 mL) was added dropwise at 0 OC with vigorous stirring to a mixture of lithium dispersion (21 mmol, 3.3% Na) in ether over a 5-min period. The reaction was allowed to warm to room temperature followed by sonication for 8 h and stirred vigorously for 5 days. The excess lithium was destroyed by careful addition of Saturated aqueous ammonium chloride solution and the mixture was extracted with ether. Purification by elution through silica gel column (hexane-ethyl acetate, %lo) followed by recrystallization from boiling cyclohexaneyielded 0.208 (42 %) of the alcohol: mp 237.5 OC; IR (CC4) Y 3348,2908,2840,1540,1260,1003,784cm-1; 'H NMR (300 MHz, THF-ds) 6 7.7-7.4, 2.0, 1.59; "C NMR (75 MHz, THF-de) 6 150.4,130.0,129.5 (2Jcp= 30.2 Hz), 128.4 ( V c p = 30.2 Hz), 125.9 ( l J c p = 301.8 Hz), 124.6 ('JCp = 3.8 Hz), 122.6 C3Jcp= 3.8 Hz), 83.9,45.5,40.1,37.8,30.2;19FNMR (188MHz, CDCl3) 6 -62.66. Anal. Calcd for C & d s O C, 75.65; H, 7.94; F, 12.81. Found: C, 75.05; H, 8.30; F, 12.42. Preparation of Carbocations. FSOaH was freshly distilled before use. A 1:l mixture of FSOsH and SOzClF was added to a suspension of the precursor alcohols in SOzClF contained in 5-mm NMR tubes at 78 OC in dry ice/acetone bath. Efficient mixing of the solution was effected using a vortex stirrer.
Acknowledgment. Support of our work by the National Institutes of Health is gratefully acknowledged. M.D.H. thanks the Department of Education for a graduate fellowship to U.S.C. We also thank Prof. Joseph Casanova (CSULA) for many helpful discussions and Dr. Robert A. Aniszfield for mass spectral data. M.D.H. wishes to thank Nikolai Hartz for help with molecular modeling calculations.
